Anti-PP2AC antibody (2038S), anti-phospho-LIMK1-Thr⁵⁰⁸/LIMK2-Thr⁵⁰⁵ (3841), and anti-phospho-AMPK-Thr¹⁷² (2531) antibodies were purchased from Cell Signaling Technology (Beverly, MA). Anti-PP2AC antibody (05-421) and metformin (1396309) was purchased from Millipore Sigma (Burlington, MA). Anti-actin (SC-477778) antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). GFP-PP2AC cDNA (cat. #RG201334) and mouse monoclonal turboGFP antibody, clone OTI2H8 (TA150041) was from OriGene (Rockville, MD). PP2A phosphatase assay kit was purchased from Upstate. Lipofectin-RNAiMAX transfection reagent (13778030), Protein G Dynabeads (1003D), Alexa Fluor 488 Phalloidin (A12379), Prolong Gold Antifade (P36941), and SYBR Green PCR master mix (4309195) were obtained from Invitrogen (Waltham, MA). FuGENE HD transfection reagent (E2311) was from Promega (Madison, WI). The enhanced chemiluminescence Western blotting detection reagents (RPN2106) were obtained from Thermo Fisher Scientific (Waltham, MA). Human α-Thrombin (HT 1002a) from Enzyme Research Laboratories (South Bend, IN). ECIS electrodes (8W10E + PET) were procured from Applied Biophysics (Troy, NY).

Human lung microvascular ECs and EGM-2MV Bullet Kit (nos. CC-3156 and CC-4147) were obtained from Lonza (Walkersville, MD). ECs were cultured in EGM-2-MV medium in 0.2% gelatin-coated flasks and maintained at 37°C in a humidified atmosphere of 5% CO2 and 95% air. ECs between passages 3 and 7 were used for experiments. GFP-PP2AC cDNA was transfected in HLMVE cells using FuGENE HD transfection reagents. Briefly, 60%-70% confluent ECs were transfected in a 6-well using 1 μg GFP-PP2AC cDNA. After 48 h of transfection, ECs were used for different experiments.

ECs at 50%–60% confluency were transfected with scramble-siRNA (sc-siRNA) and PP2AC-siRNA at a final concentration of 100 nM by using lipofectamine RNAi MAX according to the manufacturer’s instructions. Samples were collected at different time points to determine expression of target genes. After 48 h of transfection, ECs were used for different experiments. ON-TARGETplus SMARTpool-siRNA for human-PP2AC (Catalog ID:L-003598-01-0010) and ON-TARGETplus Non-targeting Control Pool (Catalog ID:L-001810-10-05) was obtained from Dharmacon.

Resistance across EC monolayers were measured by using an electrical cell substrate impedance-sensing system (ECIS; Applied Biophysics) as described previously (Siddiqui et al., 2019). ECs were seeded on gelatin coated standard array-gold plated electrodes. Cells were transfected with indicated siRNA or cDNA for 48 h. The change in TEER was measured across the monolayer in different experimental conditions and resistance was normalized to the initial value.

After 48 h of transfection, ECs were either left untreated or treated with 25 nM of thrombin in the presence and absence of metformin (10 mM for 2 h). Cells were fixed in 4% paraformaldehyde at room temperature for 10 min. After washing three times with PBS, cells were blocked with 1% BSA for 1 h at room temperature and then incubated with FITC-phalloidin (1:100 dilution) for 1 h at room temperature to stain actin stress fibers. Samples were mounted with prolong gold antifade reagent (Molecular Probes; Invitrogen) according to previous protocol (Siddiqui et al., 2019). Immunofluorescence microscopy was performed with a BZ-X710 all-in-one fluorescence microscope (Keyence, Itasca, IL). We randomly selected four to six fields per coverslip per experimental condition.

For immunoprecipitation analysis, metformin treated ECs were washed with ice-cold PBS and immediately lysed in modified radioimmunoprecipitation assay (RIPA) buffer. Equal amounts of protein (500 μg) from each sample were precleaned with control IgG for 1 h at 4°C and then incubated with anti-PP2AC (5 μg) or anti-IgG (5 μg) antibodies overnight followed by the addition of protein-G Dynabeads for 2 h at 4°C to pull down the immunocomplexes. The samples were centrifuged at 1,000 × g for 2 min and the pellet containing beads were washed three times with RIPA buffer at 4°C. After centrifugation at 1,000 × g for 2 min, the beads were collected by removing supernatant buffer, and 30 μL of 2x Laemmli buffer was added to the beads and boiled. For Western blotting assay, equal amount of proteins were separated by SDS-PAGE and probed with different primary (1:1,000 dilution, overnight at 4°C) and the corresponding secondary antibodies (1:5,000 dilution, 1 h at room temperature). Densitometry analyses was performed using ImageJ software (NIH).

Total cellular RNA was extracted from HLMVE cells using TRIzol reagent according to the manufacturer’s protocol. Reverse Transcription was performed with 1 μg of total RNA with a high-capacity cDNA reverse transcription kit from Applied Biosystems (Foster City, CA). Complementary DNA (cDNA) was then used as a template for amplification using the following primers: human IL-1β, forward, 5′-AGC​TAC​GAA​TCT​CCG​ACC​AC-3′ and reverse, 5′-CGT​TAT​CCC​ATG​TGT​CGA​AGA​A-3′; human IL-6, forward, 5′-CCT​GAA​CCT​TCC​AAA​GAT​GGC-3′ and reverse, 5′-TTC​ACC​AGG​CAA​GTC​TCC​TCA-3′; human β-actin, forward, 5′- AGC​CAT​GTA​CGT​TGC​TAT-3′ and reverse, 5′-GAT​GTC​CAC​GTC​ACA​CTT​CA-3′. The SYBR green gene expression assays were used for PCR amplification by using QuantStudio 6 Flex real-time PCR system from Applied Biosystems. Cycling conditions were as follows: 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min.

Immunoprecipitation based PP2A specific activity was analyzed by using Active PP2A DuoSet IC kit (R&D systems, MN) according to the manufacturer’s instruction. Equal amounts of cellular proteins were immunoprecipitated with an immobilized capture antibody specific for the catalytic subunit of PP2A. Ser/Thr phosphatase activity was determined by dephosphorylation of Ser/Thr phosphatase-specific phosphopeptide by active PP2A to generate free phosphate and unphosphorylated peptides. The inorganic phosphate released was detected by a sensitive dye-binding assay with malachite green and the activity of PP2A was calculated by determining the rate of phosphate released.

GraphPad Prism 20.0 software was used for the statistical analysis. Normality distribution of data was checked using Shapiro–Wilk test. For continuous variables, two tailed Student’s t-test was performed to compare experiments containing two groups. One-way ANOVA with Tukey Post-Hoc performed for experiments having more than two groups. All the experiments were repeated three independent times and the data are presented as Mean ± S.D. All p < 0.05 values were statistically significant. ^∗∗∗ denotes p < 0.001, ^∗∗ denotes p < 0.01, ^∗ denotes p < 0.05.

We first examined the effect of metformin on phosphorylation of AMPK at Threonine¹⁷². Human lung-ECs were pretreated with 10 mM of metformin for different time points. Figure 1A showed that metformin-induced highest upregulation of AMPK-phosphorylation-Thr¹⁷² residue at 2 h. Next, we studied the dose dependent effect of metformin on thrombin-induced endothelial hyperpermeability by measuring trans-endothelial electrical resistance (TEER). Human lung-ECs were pretreated with 5 and 10 mM of metformin for 2 h followed by stimulation with thrombin (25 nM). As shown in Figure 1B, thrombin challenge significantly decreased TEER values, an indicator of ECs barrier disruption and increased paracellular permeability. Intriguingly, metformin exhibited its pharmacological effects on ECs in a dose-dependent manner. The result showed that 10 mM metformin treatment most significantly protected against thrombin-induced ECs barrier disruption and increased paracellular permeability. Based on above results, we pre-treated ECs with 10 mM metformin for 2 h for further experiments. The role of F-actin stress fibers reorganization and ECs contraction in inflammation-induced paracellular gap formation is widely implicated. We next investigated the effect of metformin on thrombin-induced F-actin stress fibers formation. Our data showed a significant increase in stress fibers formation after 30 min of thrombin treatment, while pretreatment of ECs with metformin significantly decreased the level of stress fibers (Figure 1C). The upregulation of IL-6 and IL-1β is key marker of propagation of inflammation. Our qRT-PCR data showed a significant upregulation of thrombin-induced IL-6 and IL-1β mRNA levels while pretreatment of metformin significantly reduced IL-6 and IL-1β mRNA levels (Figure 1D), indicating anti-hyperpermeability and anti-inflammatory function of metformin in human lung-EC.

We next investigated the importance of upstream regulation of stress fibers formation by metformin. Ser³ phosphorylation mediated deactivation of cofilin-1 plays an important role in actin cytoskeleton reorganization and stress fibers formation in response to inflammatory insult. We hypothesized that dephosphorylation of cofilin-1 could function as a positive regulator of barrier function in response to metformin. In Figure 2, Western blot analysis showed that thrombin induced cofilin-1 phosphorylation at its inhibitory Ser³ residue. This increased in cofilin-1 phosphorylation peaked at 15 min and began to decline after thrombin challenge. Interestingly, pretreatment of ECs with metformin significantly impaired thrombin-induced cofilin-1 phosphorylation at Ser³ demonstrating that cofilin activation is involved in reducing metformin mediated stress fibers formation.

PP2A family protein phosphatases are upstream regulator of cofilin-1 that are responsible for the dephosphorylation and reactivation of cofilin-1. Therefore, we determined whether metformin regulated cofilin dephosphorylation is caused by PP2AC. To this end, we knocked down PP2AC protein in ECs using siRNA approach and assessed phospho-cofilin-1 level. A strong downregulation of the PP2AC protein was observed at 48 h post-transfection of the PP2AC-targeting siRNA duplex (Supplementary Figure S1). As shown in Figure 3A, we found that metformin significantly inhibited thrombin-induced Ser³phosphorylation of cofilin-1 in sc-siRNA ECs, however, silencing of PP2AC resulted more significant increase in phospho-cofilin-1- Ser³ levels in response to thrombin. Interestingly, dephosphorylation of cofilin-1 by metformin was markedly abrogated in PP2AC silenced ECs after thrombin challenge. We next addressed whether silencing of PP2AC would dampen metformin ability to inhibit thrombin-mediated endothelial barrier dysfunction and stress fibers formation. Our results show that siRNA mediated deletion of PP2AC markedly augmented thrombin-induced endothelial barrier dysfunction and stress fibers formation as compared to sc-siRNA ECs (Figures 3B, C) and deletion of PP2AC significantly attenuated the efficacy of metformin in preventing thrombin-induced endothelial hyper-permeability (Figure 3B) and stress fibers formation (Figure 3C). Intriguingly, phospho-cofilin and TEER data showed that deletion of PP2AC has no effect at basal level, however, thrombin challenge caused much elevation of phospho-cofilin and barrier dysfunction in ECs with PP2AC deletion suggesting an essential role of PP2AC in metformin-mediated repressing ECs dysfunction in response to thrombin. Having demonstrated a key role for PP2AC in maintaining phospho-cofilin-1-Ser³ level and in mitigating thrombin induced ECs dysfunction, we speculated that metformin could be influencing PP2AC/cofilin-1 interaction. To address these possibilities, we first studied the interaction of cofilin-1 and PP2AC by co-immunoprecipitation assay. Metformin treated cell lysates were immunoprecipitated with anti-PP2AC antibody and immunoblotted for cofilin-1. Immunoblot analysis revealed PP2AC in the pull-down preparation, while metformin treatment did not induce the interaction of PP2AC with cofilin-1 (Supplementary Figure S2 ). To investigate whether the decrease in phospho-cofilin-1 level is a result of PP2AC phosphatase and not the lack of cofilin-1 associated kinase, we therefore sought to determine the effect of metformin on LIM kinase-1 phosphorylation. In accordance with the previous data, our immunoblot results showed that thrombin induced LIMK-1 phosphorylation at Thr⁵⁰⁸ residue, however this activated phosphorylation did not differ in response to thrombin ± metformin treatment (Supplementary Figure S3). Together, these results suggested a key role for PP2AC signaling as an upstream regulator of cofilin-1-Ser³ dephosphorylation in response to metformin in lung-ECs.

We next determined the effect of metformin on Ser/Thr phosphatase activity of PP2AC. We performed an immunoprecipitation based PP2A-specific activity assay and interestingly found significant increase in PP2A activity upon ECs stimulation with metformin (Figure 4A). We then sought to determine the exact mechanism by which metformin upregulates PP2AC activity. It is established that phosphorylation at Y³⁰⁷ and methylation at Leu³⁰⁹ are the two major modifications that have been shown to modulate Ser/Thr phosphatase activity of PP2AC. It is known that Y³⁰⁷ dephosphorylation of the catalytic subunit of PP2A leads to its functional activation while methylation at Leu³⁰⁹ regulates the biogenesis of the PP2A holoenzyme. In view of these observations, we first evaluated the effect of metformin on Y³⁰⁷ phosphorylation of PP2AC. Western blot analysis with PP2AC-Y³⁰⁷ specific-antibody revealed no significant difference in the level of Y³⁰⁷ phosphorylation of PP2AC after metformin treatment (Figure 4B). We subsequently evaluated Leu³⁰⁹ methylation of PP2AC which might be responsible for increased PP2AC activity in response to metformin. Metformin treated ECs lysates were immunoblotted with C-terminal leucine residue (Leu³⁰⁹) of the PP2A catalytic subunit specific antibody. Immunoblot results showed low levels of methylated PP2AC in untreated control, while metformin supplementation significantly upregulated methylation of PP2AC in ECs (Figure 4C).